Token导航 LogoToken导航TokenDH.com
前端设计只读github未标认证来源可访问clear审计通过

learning-multilingual-assessment学习多语言评估

Agent Skill

learning-multilingual-assessment 用于记录任务执行中的错误、用户纠正、经验和能力缺口,适合在 Codex、Claude、Cursor、Gemini CLI 中希望让 Agent 持续沉淀问题、修正和最佳实践时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

220

周安装

9

GitHub Stars

1

下载量

71
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

3

许可证

MIT

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:learning-multilingual-assessment(学习多语言评估)
来源仓库:https://github.com/pauljbernard/content
仓库路径:skills/learning-multilingual-assessment
安装命令:
npx skills add https://github.com/pauljbernard/content --skill learning-multilingual-assessment
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。不同来源提供的安装方式可能略有差异;本站展示可直接复制的安装命令,安装前请核对来源页面。

skills.shnpx skills
npx skills add https://github.com/pauljbernard/content --skill learning-multilingual-assessment

简介

用于评估多语言内容的清晰度、准确性与文化适宜性。

  • 适合在国际化产品发布前的本地化质量把关。
  • 通过 GitHub 仓库安装,支持在 AI 代理中调用。
  • 需结合目标语言母语者习惯进行语义校验。
  • 应避免直译导致的语用失误或歧义。learning-multilingual-assessment 属于前端设计类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

Learning Multilingual Assessment

Design fair, valid assessments that work effectively across languages and for multilingual learners.

When to Use

  • International testing programs
  • Multilingual classroom assessments
  • ELL/ESL student assessment
  • Translated assessments
  • Global certification exams

Key Challenges

Language-Dependent Bias

Sources of Bias:

  • Complex vocabulary unnecessary for content
  • Culture-specific scenarios
  • Idioms and figurative language
  • Text-heavy questions
  • Reading speed requirements

Assessment Translation

Challenges:

  • Linguistic equivalence ≠ difficulty equivalence
  • Some concepts harder to express in certain languages
  • Test length varies by language
  • Reading time differences

Multilingual Learner Support

Considerations:

  • Content knowledge vs. language proficiency
  • Accommodations without compromising validity
  • Fair comparison across language groups

Design Principles

1. Reduce Language Load

Strategies:

  • Use simple, direct language
  • Short sentences and paragraphs
  • Visual supports (diagrams, charts, images)
  • Minimize unnecessary text
  • Concrete > abstract language
  • Active voice > passive voice

2. Avoid Cultural Bias

Review for:

  • Cultural scenarios (unfamiliar contexts)
  • Regional references (geography, events, people)
  • Socioeconomic assumptions
  • Holiday/calendar references
  • Food, sports, leisure activities

3. Universal Design

Accessibility Features:

  • Glossaries for technical terms
  • Bilingual glossaries
  • Extended time options
  • Translation tools (for instructions, not content)
  • Text-to-speech support

4. Multiple Modalities

Beyond Text:

  • Visual representations
  • Interactive elements
  • Demonstrations
  • Hands-on performance tasks
  • Oral assessment options

Translation Guidelines

Equivalence Types

Linguistic Equivalence: Word-for-word accuracy Functional Equivalence: Same meaning, different words Psychometric Equivalence: Same difficulty across languages

Translation Process

  1. Forward translation by subject expert
  2. Backward translation to verify
  3. Reconciliation of differences
  4. Pilot testing in target language
  5. Difficulty analysis and adjustment
  6. Cultural review

Validation

Field Testing:

  • Differential item functioning (DIF) analysis
  • Compare difficulty across languages
  • Identify biased items
  • Adjust or remove problematic items

Accommodations

Linguistic Supports

Allowed Accommodations:

  • ✓ Bilingual glossaries (mathematics terms)
  • ✓ Extra time
  • ✓ Simplified language instructions
  • ✓ Test directions in native language
  • ✓ Clarification of test directions

Generally Not Allowed:

  • ✗ Translation of test items (depends on purpose)
  • ✗ Side-by-side bilingual tests (for language assessments)

CLI Interface

# Design language-fair assessment
/learning.multilingual-assessment --content "math-test/" --reduce-language-load --output fair-test.md

# Validate translation equivalence
/learning.multilingual-assessment --source "test-en.json" --translations "test-es.json,test-zh.json" --validate-equivalence

# Design with accommodations
/learning.multilingual-assessment --assessment "science-exam/" --accommodations "glossary,extended-time,visual-supports"

# Cultural bias review
/learning.multilingual-assessment --test "reading-test/" --bias-check --cultures "Hispanic,East Asian,Middle Eastern"

Output

  • Language-fair assessment design
  • Translation guidelines
  • Cultural bias analysis
  • Accommodation recommendations
  • Validation protocols
  • Equivalence reports

Composition

Input from: /curriculum.assess-design, /learning.translation-quality Works with: /learning.cultural-adaptation, /learning.language-level-calibration Output to: Fair, valid multilingual assessments

Exit Codes

  • 0: Multilingual assessment designed
  • 1: Excessive language dependence
  • 2: Cultural bias detected
  • 3: Translation equivalence compromised

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

04

需要参考平台分布和安装热度时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

补充不同宿主或平台的使用分布数据

能力 5

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Claude Code

27.4%
按下载量换算19

windsurf

23.73%
按下载量换算17

zencoder

17.75%
按下载量换算13

trae

12.01%
按下载量换算9

OpenCode

8.42%
按下载量换算6

Codex

3.34%
按下载量换算2

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。

来源信息

继续浏览同类 Skills